The present pneumococcal vaccine contains the capsular polysaccharides (PS) from 23 types. In addition, there are a number of pneumococcal PS-protein conjugate vaccines under development. Antibodies to the different type PSs have been measured by a radioactive antigen binding assay, but this assay can not differentiate between antibodies to the type PS and those to cell wall contaminants found in each of the vaccine polysaccharides. We have therefore worked on standardization of an ELISA for estimation of IgG antibodies to the type PSs, and on standardization of the antibody content of a human reference serum. The published ELISA for pneumococcal antibodies, involving direct adsorption of the PS to the plastic, was compared to our published procedure that used an admixture of PS and methylated human serum albumin (mHSA). We found more uniform attachment of 12 different PS using mHSA. Since each of the type polysaccharides used in the vaccine contains a small amount of contaminating cell-wall or C PS, a way to remove the antibodies to C PS was needed. We compared different C PS absorbents and found that addition of 1 ug/mL of C PS to the serum buffer was sufficient to remove the C PS antibodies without the need for preabsorption. The FDA obtained a plasma pool from 18 pneumococcal vaccine immunized donors. This pool was aliquoted and freeze-dried for use as national reference serum for estimation of pneumococcal PS antibodies. Antibody concentrations to 9 types were provisionally assigned by scientists at Lederle-Praxis Biologicals. We evaluated the assigned IgG antibody levels by cross-standardization using our ELISA using each type to standardize the other types. We found good agreement for all except two types, types 3 and 5, and are now examining whether the mode of attachment of the PS to the plate could be responsible for the observed differences.